Serum 25(OH)D Levels and Risk of Nonalcoholic Fatty Liver Disease in Nonobese and Lean Individuals

Abstract

Purpose

The impact of vitamin D deficiency on nonalcoholic fatty liver disease (NAFLD) risk in individuals without obesity or insulin resistance has not been thoroughly evaluated. We aimed to identify whether low serum levels of 25(OH)D independently contribute to NAFLD risk in nonobese or lean individuals.

Materials and Methods

This study analyzed 241208 asymptomatic health check-up examinees who had abdominal ultrasonography. NAFLD risk was evaluated based on obesity status and serum 25(OH)D levels.

Results

The overall NAFLD prevalence was 25.5%. Among the 178630 nonobese and 126909 lean participants, the prevalence rates were 13.4% and 6.7%, respectively. The multivariable adjusted odds ratios (ORs) [95% confidence intervals (CI)] for the prevalence of NAFLD, comparing serum 25(OH)D levels of 10–19 and ≥20 ng/mL with <10 ng/mL, were 0.96 (0.93–0.99) and 0.80 (0.77–0.83), respectively. Among nonobese participants, the corresponding adjusted ORs (95% CI) were 0.94 (0.90–0.99) and 0.77 (0.73–0.81), respectively. Similar results were observed among lean participants, with those having a 25(OH)D level of ≥20 ng/mL demonstrating a significantly lower odds of NAFLD (adjusted OR, 0.76; 95% CI, 0.70–0.83). Moreover, these results were consistent even among nonobese and lean individuals who showed no signs of insulin resistance.

Conclusion

Insufficient 25(OH)D levels independently increased the risk of NAFLD, suggesting its role in the NAFLD pathogenesis, regardless of obesity or insulin resistance status. Considering the established relationship between vitamin D deficiency and nonobese/lean NAFLD, maintaining adequate 25(OH)D levels may aid in preventing the development of NAFLD, even among nonobese or lean individuals.

Graphical Abstract

INTRODUCTION

Nonalcoholic fatty liver disease (NAFLD) includes various liver conditions, from simple steatosis to cirrhosis.¹,² It is the most prevalent chronic liver condition globally, affecting approximately 1/4 of individuals.¹,² This multifaceted disease not only poses risks for liver complications, such as hepatocellular carcinoma, but also increases the likelihood of extrahepatic issues, including cardiovascular diseases and other cancers, which can lead to higher mortality rates.¹,³,⁴ Many individuals with NAFLD do not exhibit symptoms, and the prevalence is expected to grow alongside rising obesity and diabetes rates.² However, there are no approved drugs available to treat this condition, so management largely focuses on lifestyle changes, such as enhancing physical activity and adherence to a healthy diet, both of which are effective in managing initial stages of the disease and enhancing metabolic health.⁵ Notably, NAFLD can also manifest in individuals without obesity, known as nonobese or lean NAFLD.⁶ To mitigate the rising incidence of NAFLD and its associated complications, identifying and managing the adjustable risk factors linked to nonobese/lean NAFLD is very important.

Vitamin D serves multiple functions in addition to regulating calcium balance and supporting bone health.⁷ Numerous studies have emphasized the impacts of vitamin D on different areas of health, revealing an inverse correlation with conditions such as cardiovascular illnesses, diabetes, and insulin resistance.⁸ Moreover, growing evidence links a lack of vitamin D to a higher risk of NAFLD. Research in animals indicates that vitamin D is crucial for managing hepatic inflammation and fibrogenesis, as well as enhancing hepatic insulin sensitivity.⁹ Nonetheless, the effect of vitamin D deficiency on NAFLD risk in individuals who do not have obesity or insulin resistance remains to be fully investigated. Moreover, it remains uncertain whether vitamin D deficiency independently contributes to NAFLD risk, in addition to obesity or insulin resistance.

Therefore, the current study sought to identify whether low 25(OH)D level is an independent contributor for NAFLD, even among nonobese or lean individuals.

MATERIALS AND METHODS

Participant selection

The Kangbuk Samsung Health Study (KSHS) involved a group of adults aged 18 years and above who participated in thorough health assessments once or twice a year at the Total Healthcare Screening Center of Kangbuk Samsung Hospital, South Korea.¹⁰ This analysis specifically targeted a group of 419058 participants from the KSHS who received abdominal ultrasound (US) examinations between 2011 and 2016. Exclusions were made based on the following criteria (Fig. 1): 1) a history of any malignancy; 2) diagnosis of cirrhosis during abdominal US or a history of cirrhosis; 3) hepatitis B or hepatitis C positivity; 4) heavy alcohol intake; or 5) incomplete records of body mass index (BMI), fatty liver status, or serum levels of 25(OH)D. After excluding 177850 participants who fulfilled ≥1 of these criteria, the final total number of participants was 241208.

Fig. 1. Flowchart of the study population.

This study received ethical approval (KBSMC 2024-04-027). Informed consent was not necessary as our research utilized anonymized data gathered from routine health screenings in a retrospective design.

Measurements of variables

Information regarding medical history and demographics was gathered using standardized questionnaires filled out by participants. Trained personnel conducted measurements of serum biochemical parameters, blood pressure (BP), and anthropometric data during health assessments.¹⁰

Participants were grouped into three smoking statuses: never smokers, former smokers, and current smokers. Alcohol use was quantified by the number of “soju” servings, a widely consumed alcoholic drink in Korea.¹¹ A reference chart converting various alcoholic drinks into soju equivalents was used to ensure accurate reporting of alcohol intake.¹⁰ Average daily alcohol intake was determined by assessing both frequency and quantity consumed during drinking episodes. Heavy alcohol use was categorized as ≥30 g/day for men and ≥20 g/day for women.¹ Weekly physical activities were assessed, where health-enhancing physical activity (HEPA) was defined as participating in over 1500 metabolic equivalent (MET) minutes of activity per week. This included engaging in vigorous exercise for ≥3 days, or exceeding a total of 3000 MET minutes of weekly activity.¹² Hypertension was identified as having a systolic BP ≥140 mm Hg, a diastolic BP ≥90 mm Hg, a previous diagnosis of hypertension, or the current use of antihypertensive drug. Diabetes mellitus was identified by a fasting blood glucose (FBG) ≥126 mg/dL, a glycated hemoglobin ≥6.5%, a previous diagnosis of diabetes, or the current use of antidiabetic drugs or insulin.

Blood samples were collected following a minimum 10-hour fasting period for various analyses, including liver function tests, markers for viral hepatitis, lipid profiles, FBG, and insulin. Insulin resistance was evaluated using the homeostatic model assessment (HOMA-IR), which was defined by the following formula: fasting blood insulin (IU/L) multiplied by FBG (mg/dL), divided by 405.

Obesity was defined as having a BMI of ≥25 kg/m², in line with the criteria for Asians.¹³ Participants with a BMI below 25 kg/m² were classified as nonobese and further divided into two groups: lean (BMI <23 kg/m²) and overweight (BMI 23–25 kg/m²).

Serum 25(OH)D measurement

Serum vitamin D levels were evaluated by measuring total 25(OH)D, which included both 25(OH)D2 and 25(OH)D3, using a competitive immunoassay with the Elecsys Vitamin D Total assay on the Modular E170 (Roche Diagnostics, Mannheim, Germany) until April 2015, after which the method was switched to the Cobas e801 (Roche Diagnostics).¹⁴,¹⁵

Serum 25(OH)D concentrations were classified into three categories: <10, 10 to 19, and ≥20 ng/mL (equivalent to <25, 25 to 49, and ≥50 nmol/L, respectively).¹⁶ This classification aligns with the guidelines for the general healthy population, which indicate that a 25(OH)D level of 20 ng/mL or higher signifies adequate vitamin D status.¹⁷,¹⁸

Abdominal US for NAFLD assessment

The evaluation of fatty liver was performed using an US system installed a 3.5-MHz transducer after fasting for at least 10 hours. Eleven radiologists, who had no knowledge of the clinical information or the objectives of the study, conducted the US examinations. Fatty liver was diagnosed by identifying a widespread increase in fine echoes within the liver tissue, relative to the echogenicity of the spleen or kidney, along with signs of bright vessel walls and deep beam attenuation.¹⁹ The reliability of the diagnoses was determined to be substantial between different observers (kappa value of 0.74) and excellent within the same observer (kappa value of 0.94).²⁰ Participants with a history of substantial alcohol intake were excluded from this study; therefore, any cases of fatty liver identified were categorized as NAFLD.

Statistical analysis

The participants’ characteristics at baseline were categorized based on their obesity status and levels of 25(OH)D, classified as <10, 10 to 19, and ≥20 ng/mL. The primary outcome measured was the prevalence of NAFLD. To assess differences among the three BMI categories, chi-square tests and one-way analysis of variance or the Kruskal–Wallis test were applied for categorical and continuous variables, respectively. Data were reported as frequencies (%), medians (interquartile ranges), or means±standard deviations (SD), based on the distribution type. Covariates were selected based on their theoretical significance and empirical evidence from previous literature. A univariate analysis was performed to estimate the associations between potential covariates and NAFLD. Multivariate logistic regression analysis was used to determine adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for NAFLD compared with the reference group. The initial model adjustment included age and sex, with additional adjustments for confounding factors such as smoking status, physical activity, educational attainment, hypertension, diabetes, dyslipidemia, obesity, and HOMA-IR, all of which showed statistically significant results in the univariate analysis.

All p-values were calculated as two-tailed, with statistical significance threshold of p<0.05, using SPSS version 21 (IBM Corp., Armonk, NY, USA).

RESULTS

At baseline, the mean age was 37.5±9.2 years, and half of the participants (50.0%) were male. The overall NAFLD prevalence was 25.5%, and the percentages of participants with serum 25(OH)D levels of <10, 10–19, and ≥20 ng/mL were 26.0%, 53.5%, and 20.5%, respectively. Serum 25(OH)D levels exhibited a positive association with age, male sex, smoking habits, physical activity, hypertension, diabetes, dyslipidemia, and obesity (Table 1). The distribution of participants by weight category was as follows: 57.1% had normal weight, 21.8% were overweight, and 21.1% were obese. Among the 62578 and 178630 participants with and without obesity, respectively, the prevalence of serum 25(OH)D level of ≥20 ng/mL was 23.2% and 19.5%, respectively, and the corresponding prevalence of NAFLD was 60.0% and 13.4%, respectively. Among the 126909 with normal weight and 51721 who were overweight, the prevalence of 25(OH)D level of ≥20 ng/mL was 18.2% and 22.7%, respectively, and the corresponding prevalence of NAFLD was 6.7% and 29.8%, respectively. Individuals with obesity were more frequently male, current smokers, and had higher rates of hypertension, diabetes, and dyslipidemia compared to those without obesity. HOMA-IR levels were also found to be higher in obese participants compared to nonobese participants (Table 2).

Table 1. Baseline Characteristics According to 25(OH)D Levels among the 241208 Participants.

Variables	25(OH)D (ng/mL)				p-trend
	Total (n=241208)	<10 (n=62833)	10–19 (n=129028)	≥20 (n=49347)
Age (yr)	37.5±9.2	36.8±8.4	37.1±8.8	39.9±10.7	<0.001
Male sex	120719 (50.0)	23153 (36.8)	67143 (52.0)	30423 (61.7)	<0.001
Current smoker	40846 (18.9)	7165 (13.2)	22653 (19.4)	11028 (24.5)	<0.001
HEPA	38719 (16.4)	8831 (14.5)	20152 (15.9)	9736 (20.1)	<0.001
High education level*	172231 (79.7)	40813 (77.7)	94682 (80.6)	36736 (79.7)	<0.001
Hypertension†	23289 (9.7)	4705 (7.5)	12033 (9.3)	6551 (13.3)	<0.001
Diabetes‡	8022 (3.3)	1638 (2.6)	4185 (3.2)	2199 (4.5)	<0.001
Dyslipidemia	86076 (35.7)	19627 (31.2)	46654 (36.2)	19795 (40.1)	<0.001
BMI	23.1±3.4	22.6±3.4	23.2±3.4	23.4±3.2	<0.001
Obesity (BMI ≥25 kg/m2)	62578 (25.9)	13160 (20.9)	34900 (27.0)	14518 (29.4)	<0.001
NAFLD	61459 (25.5)	13384 (21.3)	34295 (26.6)	13780 (27.9)	<0.001
HOMA-IR	1.07 (0.72–1.52)	1.10 (0.72–1.64)	1.19 (0.78–1.77)	1.17 (0.77–1.76)	<0.001

BMI, body mass index; HEPA, health-enhancing physical activity; HOMA-IR, homeostasis model assessment of insulin resistance; NAFLD, nonalcoholic fatty liver disease; BP, blood pressure.

Data are expressed as means±standard deviations, medians (interquartile ranges), or numbers (percentages).

^*≥College graduate; ^†Defined as systolic BP ≥140 mm Hg, diastolic BP ≥90 mm Hg, a history of hypertension, or current use of antihypertensive medications; ^‡Defined as fasting serum glucose ≥126 mg/dL, HbA1c ≥6.5%, a history of diabetes, or current use of anti-diabetic medications.

Table 2. Baseline Characteristics According to Obesity Status among the 241208 Participants.

Variables		Total (n=241208)	Normal weight (lean) (n=126909)	Overweight (n=51721)	Obese (n=62578)	p value
Age (yr)		37.5±9.2	36.3±8.5	38.9±9.2	38.9±9.6	<0.001
Male sex		120719 (50.0)	39187 (30.9)	34753 (67.2)	46779 (74.8)	<0.001
Current smoker		40846 (18.9)	13439 (12.0)	10958 (23.6)	16449 (28.8)	<0.001
HEPA		38719 (16.4)	18411 (14.8)	9143 (18.1)	11165 (18.3)	<0.001
High education level*		172231 (79.7)	91151 (80.3)	36988 (79.9)	44092 (78.5)	<0.001
Hypertension†		23289 (9.7)	5421 (4.3)	5622 (10.9)	12246 (19.6)	<0.001
Diabetes‡		8022 (3.3)	1866 (1.5)	1859 (3.6)	4297 (6.9)	<0.001
Dyslipidemia		86076 (35.7)	26792 (21.1)	21716 (42.0)	37568 (60.0)	<0.001
NAFLD		61459 (25.5)	8547 (6.7)	15395 (29.8)	37517 (60.0)	<0.001
HOMA-IR		1.07 (0.72–1.52)	0.92 (0.62–1.31)	1.16 (0.81–1.59)	1.46 (1.05–1.89)	<0.001
25(OH)D, ng/mL						<0.001
	<10	62833 (26.0)	37774 (29.8)	11899 (23.0)	13160 (21.0)
	10–19	129028 (53.5)	66043 (52.0)	28085 (54.3)	34900 (55.8)
	≥20	49347 (20.5)	23092 (18.2)	11737 (22.7)	14518 (23.2)

HEPA, health-enhancing physical activity; HOMA-IR, homeostasis model assessment of insulin resistance; NAFLD, nonalcoholic fatty liver disease; BP, blood pressure.

Data are expressed as means±standard deviations, medians (interquartile ranges), or numbers (percentages).

^*≥College graduate; ^†Defined as systolic BP ≥140 mm Hg, diastolic BP ≥90 mm Hg, a history of hypertension, or current use of antihypertensive medications; ^‡Defined as fasting serum glucose ≥126 mg/dL, HbA1c ≥6.5%, a history of diabetes, or current use of anti-diabetic medications.

Table 3 displays the association of 25(OH)D levels with the NAFLD risk. Overall, an inverse association was observed between 25(OH)D levels and NAFLD risk. After adjustment for age, sex, and other confounding factors, the adjusted ORs (95% CI) for incident NAFLD comparing 25(OH)D levels of 10–19 and ≥20 ng/mL, with the reference level of <10 ng/mL, were 0.96 (0.93–0.99) and 0.80 (0.77–0.83), respectively.

Table 3. Risk of NAFLD by 25(OH)D Levels among the 241208 Participants.

Risk factors		Crude OR (95% CI)	p value	Age- and sex-adjusted OR (95% CI)	p value	Multivariate-adjusted OR (95% CI)	p value
Age (yr)		1.04 (1.04–1.04)	<0.001	1.04 (1.04–1.04)	<0.001	1.04 (1.04–1.04)	<0.001
Male sex		5.20 (5.09–5.31)	<0.001	1.04 (1.04–1.04)	<0.001	5.18 (5.02–5.35)	<0.001
Current smoker		2.38 (2.33–2.43)	<0.001	1.16 (1.13–1.19)	<0.001	1.08 (1.04–1.11)	<0.001
HEPA		0.93 (0.91–0.95)	<0.001	0.78 (0.75–0.80)	<0.001	0.91 (0.88–0.94)	<0.001
High education level*		1.06 (1.04–1.09)	<0.001	0.91 (0.89–0.94)	<0.001	1.13 (1.09–1.17)	<0.001
Hypertension†		3.41 (3.32–3.51)	<0.001	2.09 (2.03–2.16)	<0.001	1.21 (1.16–1.26)	<0.001
Diabetes‡		5.84 (5.57–6.12)	<0.001	3.90 (3.71–4.11)	<0.001	1.32 (1.23–1.42)	<0.001
Dyslipidemia		5.36 (5.25–5.46)	<0.001	4.23 (4.14–4.32)	<0.001	2.78 (2.71–2.85)	<0.001
Obesity (BMI ≥25 kg/m2)		9.67 (9.47–9.88)	<0.001	7.30 (7.14–7.46)	<0.001	5.89 (5.74–6.04)	<0.001
HOMA-IR		2.58 (2.55–2.61)	<0.001	2.69 (2.66–2.73)	<0.001	2.34 (2.31–2.38)	<0.001
25(OH)D, ng/mL		1.01 (1.01–1.02)	<0.001	0.99 (0.99–0.99)	<0.001	0.99 (0.98–0.99)	<0.001
	<10	1 (Reference)		1 (Reference)		1 (Reference)
	10–19	1.34 (1.31–1.37)	<0.001	1.04 (1.02–1.07)	0.001	0.96 (0.93–0.99)	0.006
	≥20	1.41 (1.39–1.47)	<0.001	0.86 (0.83–0.88)	<0.001	0.80 (0.77–0.83)	<0.001

BMI, body mass index; CI, confidence interval; HEPA, health-enhancing physical activity; HOMA-IR, homeostasis model assessment of insulin resistance; NAFLD, nonalcoholic fatty liver disease; OR, odds ratio; BP, blood pressure.

The multivariable-adjusted model was adjusted for confounders including age, sex, smoking status, physical activity, educational level, hypertension, diabetes mellitus, dyslipidemia, obesity, and HOMA-IR.

^*≥College graduate; ^†Defined as systolic BP ≥140 mm Hg, diastolic BP ≥90 mm Hg, a history of hypertension, or current use of antihypertensive medications; ^‡Defined as fasting serum glucose ≥126 mg/dL, HbA1c ≥6.5%, a history of diabetes, or current use of anti-diabetic medications.

The analysis of 25(OH)D categories also revealed an inverse relationship between 25(OH)D levels and NAFLD in a dose-dependent manner in both groups with and without obesity after multivariable adjustment. The adjusted ORs (95% CI) for incident NAFLD among participants with obesity, when comparing 25(OH)D level of 10–19 and ≥20 ng/mL with the reference level of <10 ng/mL, were 0.92 (0.87–0.97) and 0.78 (0.74–0.83), respectively. Among nonobese participants, the corresponding adjusted ORs (95% CI) were 0.94 (0.90–0.99) and 0.77 (0.73–0.81), respectively. Similar results were also observed among lean participants, with those having a 25(OH)D level of ≥20 ng/mL demonstrating a significantly decreased NAFLD risk (adjusted OR, 0.76; 95% CI, 0.70–0.83) (Fig. 2, Supplementary Tables 1 and 2, only online).

Fig. 2. Risk of NAFLD by 25(OH)D levels among nonobese (n=178630) and lean participants (n=126909) with (A) and without (B) insulin resistance. NAFLD, nonalcoholic fatty liver disease; OR, odds ratio; CI, confidence interval.

In nonobese and lean individuals without insulin resistance, indicated by a HOMA-IR value of less than 2.5, the NAFLD risk in relation to the 25(OH)D categories was consistent, with higher 25(OH)D levels being associated with a significantly decreased NAFLD risk (Fig. 2, Supplementary Tables 1 and 2, only online).

We further analyzed the data stratified by age and sex (Supplementary Tables 3, 4, 5, only online). The results were consistent with our original findings, but the association of low 25(OH)D levels with the NAFLD risk was more prominent in individuals under the age of 50 years and in males.

DISCUSSION

In this study investigating the association of serum 25(OH)D levels with the NAFLD risk in a large cohort of asymptomatic examinees, the 25(OH)D levels were significantly and inversely associated with the NAFLD risk in a dose-dependent manner. Interestingly, the protective relationship between higher 25(OH)D levels and reduced NAFLD prevalence was consistent even among nonobese individuals. In particular, higher 25(OH)D levels significantly decreased the risk of NAFLD, even in lean individuals. Moreover, these results were consistent in nonobese or lean individuals without insulin resistance. Insufficient 25(OH)D level was an independent contributor for NAFLD, suggesting that 25(OH)D levels might significantly influence the pathogenesis of NAFLD, irrespective of obesity and insulin resistance status.

Accumulating evidence suggests a significant association of vitamin D deficiency with an increased risk of type 2 diabetes and cardiovascular diseases, all of which are commonly related to the development of NAFLD.²¹,²²,²³ Despite some discrepancies, a large amount of epidemiological data supports that low 25(OH)D levels are substantially linked to an increased NAFLD risk.⁵,²⁴,²⁵,²⁶ A cohort study from Korea has reported an inverse relationship between serum 25(OH)D levels and NAFLD development, as well as a positive correlation with NAFLD resolution.⁵ A recent meta-analysis, encompassing 15 high-quality studies, has also demonstrated that individuals with NAFLD have 25(OH)D levels that are 0.90 ng/mL lower compared to those without NAFLD, and the 25(OH)D level is negatively correlated with the risk of NAFLD (OR, 0.64; 95% CI, 0.54–0.77).²⁴ In a cross-sectional analysis utilizing data from the Third National Health and Nutrition Examination Survey of United States with a median follow-up of >18 years, low vitamin D level was significantly correlated not only with the severity of NAFLD and liver fibrosis, but also with an increased mortality rate in NAFLD, especially in relation to diabetes.²⁷ Another cross-sectional study comparing 60 patients with biopsy-confirmed NAFLD and 60 healthy subjects has demonstrated that patients with biopsy-confirmed NAFLD had a significantly increased hypovitaminosis D3 prevalence and notably lower 25(OH)D levels compared to matched controls.²⁸ Moreover, the severity of NAFLD, as determined by histological examination, can be anticipated based on the levels of 25(OH)D.²⁸ Individuals diagnosed with NASH showed reduced levels of 25(OH)D compared to those with simple steatosis.²⁸ Consistent with these studies, our multivariate analysis, which was adjusted for statistically significant potential variables identified in both previous studies and our data, confirmed an independent inverse correlation between 25(OH)D levels and NAFLD risk in a dose-dependent manner. Moreover, one Mendelian randomization study involving three European-descendent populations has reported an independent protective impact of 25(OH)D levels on the development of NAFLD, demonstrating a concomitant decrease in the risk of NAFLD (OR, 0.78; 95% CI, 0.69–0.89) for each 1-SD increase in genetically predicted 25(OH)D levels.²⁶

While NAFLD is commonly related to obesity, it can also manifest in nonobese individuals. A meta-analysis has revealed that approximately 40.8% of individuals with NAFLD are classified as nonobese and 19.2% as lean.²⁹ In agreement with these findings, among the participants with NAFLD in our study, 39.0% were nonobese and 13.9% were classified as lean. Nonetheless, the effect of the 25(OH)D status on the risk of nonobese/lean NAFLD, especially in cases without insulin resistance, has not been thoroughly investigated. Only one study has examined this issue to date. A cross-sectional study involving 613 nonobese individuals with BMI of <30 kg/m² demonstrated that a low 25(OH)D level was a significant determinant of NAFLD.⁸ However, this study had some limitations, including a relatively small sample size, an older and heterogeneous patient group with a mean age of >50 years, potentially diverse comorbidities, and a BMI cutoff value for nonobesity that was not specific to Asians. To the best of our knowledge, our investigation, derived from a larger and more homogeneous cohort primarily comprising young and middle-aged asymptomatic individuals, is the first to reveal a consistently significant association of vitamin D deficiency with the risk of NAFLD, even among nonobese and lean Asians and those without insulin resistance. While previous studies have suggested that the association between vitamin D deficiency and NAFLD may be primarily mediated by insulin resistance, our study provides new insights by stratifying the data according to obesity and HOMA-IR status. These findings indicate that keeping sufficient levels of 25(OH)D is helpful in lowering the risk of NAFLD, even among nonobese or lean individuals. This highlights the importance of vitamin D in the context of NAFLD, extending beyond the traditional factors of obesity or insulin resistance.

Although the specific biological mechanisms connecting vitamin D deficiency and NAFLD in nonobese or lean individuals remain unclear, there are several potential explanations that extend beyond obesity and insulin resistance.³⁰ 25(OH)D exhibits systemic and hepatic tissue-specific anti-inflammatory, antifibrotic, and immunomodulatory properties, contributing to the association. In various animal studies, low 25(OH)D levels exacerbated hepatic steatosis, lobular inflammation, and fibrosis by increasing oxidative stress, enhancing proinflammatory substances such as tumor necrosis factor-α and interleukin-6, and upregulating profibrotic mediators including platelet-derived growth factor collagen, while decreasing toll-like receptors and adiponectin secretion.⁷,⁹,³⁰,³¹,³² Human studies have also revealed that hepatic vitamin D receptor expression inversely correlates with lobular inflammation and steatosis severity in liver histology.³³ Additional studies are required to evaluate the possible advantages of vitamin D supplementation for preventing and treating NAFLD in nonobese or lean individuals.

This study has some limitations. First, we lacked information on factors that could affect serum 25(OH)D levels, including dietary vitamin D; specifics about the quantity, frequency, and type of vitamin D supplements; sunlight exposure; outdoor activities; or the existence of genetic polymorphisms. As a result, there is still a chance of residual confounding that might influence the results of the association. Second, the findings of this study were obtained from a cohort of relatively young and healthy participants compared to participants in many other studies on NAFLD. Consequently, these results might not be applicable to older or ethnically diverse groups. However, considering the rapid increase in vitamin D deficiency among young adults³⁴,³⁵ and the potential for fewer study participants with associated comorbidities linked to vitamin D deficiency, our study cohort, predominantly composed of relatively young and asymptomatic participants, may provide a more accurate estimate for the impact of vitamin D on the risk of nonobese/lean NAFLD without insulin resistance. Lastly, the study’s cross-sectional design hampered the ability to identify temporal or causal relationships. While vitamin D levels and obesity can vary over time, assessing the effects of these dynamic changes on NAFLD risk was not possible since most examinees were assessed only once during their first visit. Nevertheless, the substantial sample size may help reduce potential biases associated with these variations.

In conclusion, serum 25(OH)D levels have a negative correlation with the NAFLD risk. The protective relationship between higher 25(OH)D levels and reduced NAFLD prevalence was consistently observed even in nonobese or lean individuals without insulin resistance. Our results suggest that vitamin D deficiency is an independent contributor for NAFLD, and that 25(OH)D may play a substantial contribution in the NAFLD pathogenesis, regardless of obesity and insulin resistance status. Considering the clear connection between vitamin D deficiency and nonobese/lean NAFLD, maintaining adequate 25(OH)D levels may aid in preventing the development of NAFLD, even among nonobese or lean individuals. Additional prospective studies are necessary to explore whether vitamin D supplementation can promote the regression of NAFLD or decrease its incidence and related complications.

SUPPLEMENTARY MATERIALS

Supplementary Table 1

Risk of NAFLD by 25(OH)D Levels among Nonobese (n=178630) and Lean Participants (n=126909)

Supplementary Table 2

Risk of NAFLD by 25(OH)D Levels among Nonobese (n=168211) and Lean Participants (n=121761) without Insulin Resistance^*

Supplementary Table 3

Risk of NAFLD by 25(OH)D Levels According to Age and Sex

Supplementary Table 4

Risk of NAFLD by 25(OH)D Levels among Nonobese and Lean Participants According to Sex and Age

Supplementary Table 5

Risk of NAFLD by 25(OH)D Levels among Nonobese and Lean Participants without Insulin Resistancea According to Sex and Age